Hide/Show Apps

Future of green building contracts in Turkey

Mohammadi, Sahra
Conceptual cost estimate is very important for initial project decisions when the design information is limited and the scope is not finalized at the early stages of the construction projects. It has serious effects on planning, design, cost management and budgeting. Therefore, the decision makers should be as accurate as possible while estimating the conceptual cost at the initial stage since a misestimation on the conceptual cost may lead to serious problems during feasibility analysis or at the later stages of the projects. In this thesis, a support vector regression method is presented in order to estimate the conceptual cost of construction projects. For this purpose, 10 historical cost data sets including 273 projects were compiled and analyzed by the proposed method. The proposed method enables identification of parsimonious mapping function between the independent variables and the cost. Besides, it presents a robust and pragmatic alternative for conceptual cost estimation of construction projects. The results of the analyses by the proposed method were also compared with the estimates obtained by two other machine learning methods, which are neural network and case based reasoning, in terms of their prediction accuracy. The results indicate that the proposed method outperforms existing state-of-art machine learning methods for conceptual cost estimation of construction projects.